Such a method and such an arrangement are known from U.S. Pat. No. 5,692,472. There, to control the drive unit, the torque or the power of the drive unit is adjusted electrically at least in dependence on the position of an operator-controlled element actuated by the driver. A maximum permissible torque or a maximum permissible power is determined on the basis of the position of the operator-controlled element as well as at least on the engine rpm. The maximum permissible torque or the maximum permissible power of the drive unit should not exceed the torque or the power during the actual operating state. The actually adjusted torque or the actually adjusted power of the drive unit is determined from operating variables such as engine rpm and the inducted air mass. This actual adjusted torque or power is compared to the maximum permissible value and a fault reaction is initiated when the computed torque or the computed power exceeds the maximum permissible torque or the maximum permissible power. With this monitoring measure, the operating reliability of the drive unit is ensured because a torque generation of the drive unit, which is increased compared to the driver command, can be prevented in a reliable manner. The response of the shown monitoring is only wanted in the actual case of a fault. In addition, operating situations are conceivable (for example, in transition states), in which the monitoring responds to tightly pregiven tolerances without a fault being present. Such a behavior is not wanted.